Transcript The Astrophysical Origins of the Short

The Astrophysical Origins of the
Short-Lived Radionuclides in the
Early Solar System
Steve Desch
September 15, 2006
U. of Toronto
with a shout-out to my ASU supernova posse:
Jeff Hester, Nicolas Ouellette, Carola Ellinger
Outline
• Short-lived radionuclides:
– What are they?
– How are they measured?
• Possible sources:
– Inheritance
– Irradiation
– Injection
• “Aerogel” model:
– Astrophysical context
– SLR predictions
Short-Lived Radionuclides
“SLRs” = Radionuclides with
half-lives t1/2 < 16 Myr
Early Solar System SLRs Confirmed
by Isotopic Analyses of Meteorites:
41Ca
(t1/2 = 0.1 Myr) (Srinivasan et al. 1994, 1996)
36Cl (t
1/2 = 0.3 Myr) (Murty et al. 1997; Lin et al. 2004)
26Al (t
1/2 = 0.7 Myr) (Lee et al. 1976)
60Fe (t
1/2 = 1.5 Myr) (Tachibana & Huss 2003; Mostefaoui et al. 2004)
10Be (t
1/2 = 1.5 Myr) (McKeegan et al. 2000; Sugiura et al. 2001)
53Mn (t
1/2 = 3.7 Myr) (Birck & Allegre 1985)
107Pd (t
1/2 = 6.5 Myr) (Kelly & Wasserburg 1978)
182Hf (t
Myr) (Harper & Jacobsen 1994)
1/2 = 9
129I (t
1/2 = 15.7 Myr) (Jeffery & Reynolds 1961)
Isotopic analyses of meteorites
show they once held SLRs:
Excess 10B is from
decay of 10Be
Slope gives original
10Be/9Be ratio
“Natural”
10B / 11B ratio
McKeegan et al. (2000)
Initial Abundances of Confirmed SLRs:
Possibly 60Fe/56Fe = 1.6x10-6
irons
Unconfirmed SLRs:
7Be
(t1/2 = 57 days)
(Chaussidon et al. 2004)
63Ni
(t1/2 = 101 years) (Luck et al. 2003)
97Tc
(t1/2 = 2.6 Myr) (Yin& Jacobsen 1998)
99Tc
(t1/2 = 0.21 Myr) (Yin et al. 1992)
135Cs
(t1/2 = 2.3 Myr) (McCulloch & Wasserburg
1978; Hidaka et al. 2001)
205Pb
(t1/2 = 15 Myr) (Chen & Wasserburg 1981)
Chaussidon et al (2006)
Luck et al (2003)
Inheritance
Sun and Protoplanetary Disk may have inherited
SLRs as a result of Galactic processes:
Ongoing Galactic Nucleosynthesis
Supernovae, Wolf-Rayet winds, novae, etc., eject
newly created radionuclides into Galaxy
Galactic Cosmic Rays
Proton, alpha particle Galactic Cosmic Rays
(GCRs) spall ambient nuclei, producing SLRs
Some GCR nuclei are SLRs, get trapped in gas
that forms Solar System (Clayton & Jin 1995)
Ongoing Galactic Nucleosynthesis?
supernova
Stars form in the
spiral arms of
spiral galaxies
M 109
supernovae (and
Wolf-Rayet winds)
eject radionuclides
radionuclide-laden
gas orbits Galaxy
for ~100 Myr, until
next spiral arm
new stars form
with radionuclides
182Hf
129I
26Al
53Mn
60Fe
Harper (1996)
More sophisticated
mixing models show
41Ca, 26Al, 60Fe
definitely
not inherited from ISM.
Predicted ratios
if only sources
are type II SNe.
Jacobsen (2005)
Ongoing Galactic Nucleosynthesis
•Could explain abundance of 129I and longer-lived radionuclides with ~100 Myr delay consistent with Galactic
dynamics.
•Definitely does not explain 41Ca, 26Al or 60Fe
abundances [Harper (1996); Wasserburg et al. (1996); Meyer &
Clayton (2000); Jacobsen (2005)].
•If the majority of 60Fe really was due to ongoing
Galactic nucleosynthesis, 53Mn, 107Pd, 182Hf and 129I
would be vastly overproduced.
Galactic Cosmic Rays
•Most GCRs are protons; other nuclei present in near-solar
proportions
•Spacecraft have accurately measured fluxes of GCRs of
different nuclei and energies (10 MeV/n to > 10 GeV/n)
•Beryllium GCRs 106 times more abundant than expected from
solar abundances (i.e., 1 in 103 instead of 1 in 109).
•Flux of 10Be GCRs is known and is large
•Fluxes of all GCRs scale linearly with star formation rate,
which was almost certainly a factor of 2 higher 4.5 Gyr ago
Galactic Cosmic Rays
Galactic Cosmic Rays (GCRs)
follow magnetic field lines
Magnetic field lines observed to
converge in star-forming cores
Schleuning (1998)
GCRs funneled into cloud cores
Some GCRs mirrored
out of cloud core by B
fields
B fields funnel some
GCRs into cloud core
GCRs in cloud core
can be trapped if
column density ∑ is
high enough
Cloud core B, ∑
taken from Desch &
Mouschovias (2001)
Column Density ∑(t), Magnetic Field Strength B(t) calculated
(Desch & Mouschovias 2001; Desch, Connolly & Srinivasan 2004)
GCRs ionize gas passing through cloud core, lose energy, slow
down (Bethe formula)
Low-energy (< 100 MeV/n) 10Be GCRs are trapped when
∑ ~ 0.01 g cm-2
Desch, Connolly & Srinivasan (2004)
total 10Be/9Be
10Be
GCRs
trapped in
cloud core
10Be/9Be
in
meteorites
GCR protons spall
local CNO nuclei,
produce 10Be
Galactic Cosmic Rays
•10Be in meteorites entirely attributable to trapped 10Be GCRs
•Biggest uncertainty is GCR flux 4.5 Gyr ago (factor of 2);
probably all but at least half of 10Be is trapped GCRs
•Trapped GCRs do not explain any other SLR, but 10Be is
known to be decoupled from other SLRs (Marhas et al. 2002)
Inheritance –– Conclusions
•At least half, and probably all, 10Be is inherited
•129I may be inherited
•Other SLRs, esp. 41Ca, 26Al and 60Fe, are not inherited.
Irradiation
Energetic particles (accelerated by solar flares within
the Solar System) may have irradiated material,
inducing nuclear reactions and creating SLRs
Solar flares accelerate p, 4He, 3He to E > 10 MeV/n
Particle fluxes ~105 times larger around T Tauri stars; in
1 Myr, 1048 (!) energetic particles emitted
Irradiation within the Disk
Gas and dust in the protoplanetary disk (~ 1 AU)
Irradiation within the Sun’s Magnetosphere
Solids only, inside ~ 0.1 AU
Irradiation in the Disk
If gas is present, energetic particles lose > 99% of their energy
ionizing gas, not inducing nuclear reactions (Nath & Biermann 1994)
Consider 26Al:
26Al
/ 27Al = 5 x 10-5 implies 1045
26Al
atoms in a 0.01 M disk
Only 1048 particles emitted in 1 Myr; only 1047 intercept disk
To make a 26Al atom by 26Mg(p,n)26Al, a proton must travel
through ∑ ~ 1.4 mH / (xMg26 ) > 3 x 106 g cm-2 of gas
But protons stopped by << 10 g cm-2 of gas (Bethe formula):
fewer than 1 proton in 105 reacts (Clayton & Jin 1995)
Even including other energetic particles, other targets, can’t make
more than ~ 1042 26Al atoms
Similar results for other SLRs, including 10Be
Irradiation inside the Sun’s Magnetosphere
e.g., “X-wind” model
Shu et al. (2001)
very little gas -- it’s ionized and
part of the corona
only solids (CAIs)
are irradiated
a fraction of
the solids are
returned to
asteroid belt
Six problems with the X-wind model:
1. Launching of solids from 0.1 AU to asteroid belt problematic:
winds probably launched from 1 AU, not 0.1 AU [Coffey et al.
(2004)]; trajectories very sensitive to particle size [Shu et al. (1996)]
2. CAIs formed in near-solar f O2, but “reconnection ring” is >104
times more oxidizing than solar [using values in Shu et al. (2001)]
3. Concordant production of 26Al, 41Ca requires Fe,Mg silicate
mantle to surround Ca,Al-rich core, but real minerals do not
separate this way (e.g., Simon et al. 2002)
4. Production of 26Al or 41Ca at meteoritic levels will overproduce
10Be, using best-case scenario [Gounelle et al. (2001)] and new
measured reaction rate for 3He(24Mg,p)26Al [Fitoussi et al. (2004)],
especially if most 10Be is inherited [Desch et al. (2004)]. [See also
Marhas & Goswami (2004)]
Six problems with the X-wind model (continued):
5. Temperatures inside magnetosphere at least 750 K, and usually
> 1200 K [Shu et al. (1996)]. Chlorine (including 36Cl) requires
T < 970 K to condense [Lodders (2003)]
6. Many other SLRs cannot be produced by spallation, including
60Fe, 107Pd and 182Hf [Gounelle et al. (2001); Leya et al. (2003)]
Many of these problems pertain to any model of
irradiation in the Sun’s magnetosphere
Irradiation –– Conclusions
•Energetic-particle irradiation occurs and can produce
7Be, 10Be, 41Ca, 26Al, 53Mn, if irradiation occurs in Sun’s
magnetosphere (to minimize ionization energy losses)
•Confirmation of 7Be would demand irradiation
•Concordant production of 41Ca, 26Al difficult, 10Be
probably overproduced, and 36Cl hard to condense
•60Fe, 107Pd, 182Hf (and 36Cl?) demand external source
Injection
Stellar nucleosynthesis products ejected by an evolved
star and enter the Solar System material shortly before,
or soon after, Solar System formation:
AGB star
Contaminates Sun’s molecular cloud [wind possibly
triggers collapse of cloud core] (Wasserburg et al. 1994)
Nearby (Type II) Supernova
Contaminates Sun’s molecular cloud core and triggers
its collapse (Cameron & Truran 1977) ... or ....
Injects into already-formed protoplanetary disk...
AGB Star
Stars at least as
massive as the Sun at
the ends of their lives
enter AsymptoticGiant Branch stage
Eskimo nebula: after AGB
winds expose white dwarf
SLRs created within star are dredged up
to the surface and ejected in a powerful
wind
Problems with the AGB Scenario:
1. AGB stars do produce 41Ca, 36Cl, 26Al, 60Fe, 107Pd, 135Cs and 205Pb
[Wasserburg et al. 1994, 1995, 1996, 1998; Gallino et al. 1998, 2004].
But they do not produce 129I, 53Mn, or 182Hf.
2. AGB stars are extremely unlikely to be associated with the early
Solar System. Kastner & Myers (1994) conservatively calculate
probability of contamination of Sun’s molecular cloud core at
< 3 x 10-6
Supernovae
•Supernovae do produce all the confirmed SLRs: 41Ca, 36Cl, 26Al,
53Mn, 60Fe, 107Pd, 182Hf, 129I.
(Except for 10Be, which is
known to have a separate
origin.)
•Relative abundances of
SLRs in outermost ~18 M
of a 25 M supernova match
meteoritic values very well
[Meyer et al. 2003]
•Order-of-magnitude
agreement sufficient,
considering real supernova
ejecta highly heterogeneous
Cassiopeia A supernova remnant
time delay
= 0.9 Myr
Meyer et al (2003), LPSC abstract
Supernova and Star Formation
•Meteoritic values require Solar System disk to be 0.01% SN ejecta
•Requires supernova < 10 pc away, ~ 1 Myr before CAIs formed (see
Fields et al. 2007)
•What are the odds our Solar System “happened” be near supernova?
Like case of AGB star: too low. There must be a causal connection.
•One way in which SN could be causally connected is if the SN shock
triggered the collapse of our cloud core [Cameron (1963), Cameron & Truran
(1977)]: “supernova trigger” model
Supernova shock
can inject right
amounts of
SLRs, and
trigger collapse
of cloud core if...
Supernova shock
can be slowed to
20 - 50 km/s
Vanhala & Boss (2002)
Requires a lot of
intervening gas,
but travel times
t ~ 105 yr
Are these conditions met? Preceding state must
include H II region!
low-density, ionized gas
dense molecular gas
n ~ 104 cm-3
shocked gas
cloud
core
n ~ 10 cm-3
UV photons
ionization front
shock
= sharp density
discontinuity
~ 0.2 pc
supernova
progenitor
supernova
cloud
core
ejecta
ionization front
= sharp density
discontinuity
ejecta
cloud
core
Vej ~ 2000 km/s
Ejecta transfers its
momentum: shock
propagates to cloud core,
slowed to < 20 km/s
cloud
core
The actual ejecta (and
SLRs) do not penetrate
into cloud: they bounce!
(Hester et al. 1994)
Does this gas contain
any radioactivities?
Injection –– Conclusions so far...
•Injection by AGB stars highly unlikely, and cannot
explain all isotopes anyway (esp. 53Mn, 182Hf)
•Injection by supernovae explains all isotopes well, but
causal link to Solar System formation must be explained
•Supernova trigger viable, but needed conditions may
not exist where supernovae happen
•Alternative supernova scenario...
“Aerogel” Model
Very close (< 1 pc) supernova injected SLRs into the Solar
System, after it had formed a disk (Gold 1977; Clayton 1977;
Chevalier 2000; Ouellette et al 2005)
1 Ori C: 40 M O6 star;
will supernova in 1-2 Myr
Protostars
with disks
Orion Nebula
When 1 Ori C goes supernova, all the disks in the Orion Nebula
will be pelted with radioactive ejecta
Same scenario even more likely for disks observed in Carina
Nebula, with sixty O stars [Smith et al. (2003)], or NGC 6611
[Oliveira et al. 2005] or NGC 6357 [Healy et al. 2007, in prep]
Ejecta dust grains penetrate disk, evaporate on entry, but leave
SLRs lodged in disk like aerogel: “Aerogel Model”
Initial abundance of 26Al (26Al/27Al = 5 x 10-5) is explained
by homogeneous injection of 5 x 10-6 M of a 25 M
supernova’s ejecta into a minimum-mass (0.01 M) disk.
5 x 10-6 M is the ejecta mass intercepted by a 40 AUradius disk 0.2 pc from a 25 M supernova
But will a disk this close survive? Will ejecta be mixed in?
To answer these questions, we have written a 2-D hydro code
based on the Zeus algorithms.
Includes tensor artificial viscosity and a cooling term.
Canonical Simulation
• Disk
– Minimum mass (0.01 M) disk truncated at 30 AU
– Disk allowed to dynamically relax for 1000 years
– Final radius ~ 40 AU
• Supernova
– 0.3 pc away
– 1051 ergs (1 foe) explosion energy
– 20 M ejected isotropically with time dependence of density and velocity
consistent with Matzner & McKee (1999) and uniform-density star*
– Isotopic composition assumed homogeneous, that of 25 M supernova
from Woosley & Weaver (1995)
Relaxed Disk
Canonical Run
QuickTime™ and a
YUV420 codec decompressor
are needed to see this picture.
Reverse Shock
Disk Stripping
Stripping and Mixing: KH Rolls
Distance
Injection efficiency
26Al/27Al
0.1 pc
2.4%
7 x 10-6
0.3 pc (canonical)
0.7%
7 x 10-8
0.5 pc
0.5%
1 x 10-8
Energy
0.25 foe
Injection efficiency
0.9%
26Al/27Al
1 foe (canonical)
4 foe
0.7%
0.6%
7 x 10-8
6 x 10-8
Disk mass
Injection efficiency
26Al/27Al
0.1 x min. mass
1.2%
1.2 x 10-6
Min. mass (canonical)
0.7%
7 x 10-8
10 x min. mass
0.8%
8 x 10-9
9 x 10-8
Aerogel Model: Conclusions
• Protoplanetary disks will survive nearby supernova
explosions
• Gas-phase supernova ejecta is mixed into the disk, but with
low efficiency (~ 1%), too low to explain SLR ratios
• Dust injection is the best candidate for SLR injection and
will be the subject of future work
– Preliminary calculations show the dust will travel roughly 100 AU
before being deviated by the bow shock, and will be mixed in with ~
100% efficiency
• All SLRs inferred from meteorites were in solid phase...
Allowing supernova ejecta to be injected inhomogeneously
allows an almost perfect match to meteoritic abundances
Conclusions
•Inheritance: 10Be likely inherited (trapped cosmic rays),
129I may be inherited, but no others, especially not 60Fe!
•Irradiation: would be necessary for 7Be, but overproduces
10Be, can’t explain 182Hf, 107Pd, (36Cl?), and especially 60Fe!
•Injection: AGB star can’t explain 53Mn, 182Hf, and is very
unlikely; supernova can explain all SLRs if link to Solar
System formation made; supernova trigger viable but may
not pertain to real supernova environments
•Aerogel Model: Inevitable in supernova environments; at
first cut is consistent with data. Refinements underway!